AI / ML Engineer Bootcamp — Infrabuild Partners

01 · What this is

Python for ML, scikit-learn, deep learning in PyTorch, NLP, modern LLMs, vector databases and RAG, agents, and the MLOps that gets a model off a laptop and into production.

A live cohort, taught in real timeLive cohort · 3 evenings a week · US-Eastern. You ask your question in the room and get the answer while it still matters.

The full curriculum, module by moduleTen modules built and taught by engineers who do this work — the syllabus below is all of it, with nothing held back for an upsell.

InfraBuild Labs platform accessYour seat includes — of access to the InfraBuild Labs platform. Its interactive lab content is DevOps-first today; this track is delivered as a live cohort plus the full curriculum.

Cohort fee

—

One payment for the seat, with — of InfraBuild Labs platform access included.

Enrol now → Questions first? Talk to us →

Checkout runs on the InfraBuild Labs platform. Every application is reviewed by hand, and your seat is confirmed once it is approved.

02 · Curriculum

The whole syllabus, up front

Six weeks, ten modules, from the fundamentals that keep you honest to the LLM systems people are hiring for right now.

10 modules · 6 weeks · taught live

01 ML Foundations

Supervised vs unsupervised vs reinforcement, train/validation/test splits, the bias-variance tradeoff.

02 Math for ML

Linear algebra essentials, a calculus refresher, probability and statistics, gradients and optimisation.

03 scikit-learn

Pipelines, transformers, estimators, linear and tree models, metrics, cross-validation, tuning.

04 Deep Learning with PyTorch

Tensors, autograd, network basics, training loops, CNNs, RNNs, transfer learning.

05 NLP Fundamentals

Tokenisation, embeddings, word2vec, transformers, attention, BERT, fine-tuning Hugging Face models.

06 Large Language Models

GPT and Claude APIs, prompt design, function calling, structured outputs, evaluation, cost control.

07 RAG & Vector Databases

Embeddings, chunking strategies, Pinecone / Weaviate / pgvector, retrieval quality, hybrid search.

08 Prompt Engineering & Agents

System prompts, few-shot, chain-of-thought, ReAct, tool use, multi-agent patterns, evals.

09 MLOps

Model versioning, experiment tracking with W&B and MLflow, registries, deployment, monitoring.

10 Production AI Apps

One AI application end to end: backend, vector store, LLM, frontend, deployment, observability.

03 · Tool stack

What you will actually touch

The tools named in the job descriptions you are aiming at — learned in the order they turn up in real work.

Python NumPy Pandas scikit-learn PyTorch Hugging Face OpenAI Anthropic LangChain Weights & Biases

04 · Who this is for

Four kinds of people end up in this room

If one of these sounds like you, you will be in the right cohort. If none of them do, tell us and we will point you at the track that fits.

Python engineers

Add ML and AI to your toolkit — the highest-leverage skill of the decade.

Data scientists

Move out of notebooks and into production ML and LLM systems.

Career switchers

Come from a quantitative background — maths, physics, stats — into AI/ML.

Builders

Ship real LLM-powered products fast.

05 · Outcomes

What you will be able to do

Not topics covered — things you can do on your own once the six weeks are behind you.

Train a scikit-learn classifier with proper cross-validation.
Fine-tune a Hugging Face transformer on a custom dataset.
Build a RAG application end to end with embeddings and an LLM.
Deploy an ML model behind a FastAPI endpoint.
Build an LLM agent with tool use and structured outputs.

06 · Careers

Roles this leads to

The job titles this curriculum is built against. We teach to the work these roles do, not to a certificate.

ML Engineer AI Engineer MLOps Engineer Data Scientist LLM Application Engineer

07 · Questions

Before you enrol

The things people ask us on the call, answered here so you do not have to book one first.

Do I need experience before I start?

This is the one track pitched at Intermediate, and it matters. You should already write Python comfortably — functions, packages, a virtual environment, reading a traceback without panic. The mathematics is covered in Module 02, so linear algebra, probability and gradients are refreshed rather than assumed.

When does the next cohort run?

Next batch — enrolling now. It runs for 6 weeks — Live cohort · 3 evenings a week · US-Eastern. We confirm your exact start date when your application is approved.

Are the classes live, or recorded video?

Live. The schedule above is real class time with an instructor in the room — you interrupt, you ask, you get an answer. That is the whole reason this is sold as a cohort and not as a video library.

How much of this is hands-on?

Every module is worked through with the instructor, tool by tool, and the outcomes above are what you should be able to do unaided by the end. To be exact about what a seat buys: a live cohort, the full curriculum, and access to the InfraBuild Labs platform. The platform's interactive lab content is DevOps-first today — we would rather say so here than let you find out later.

What is InfraBuild Labs, and how long do I keep it?

It is our own lab platform: browser terminals wired to real machines, guided exercises and per-task verification. Its interactive content is DevOps-first today, so on this track treat it as the environment you practise and build in rather than a second copy of this syllabus. Your seat includes — of access.

How does enrolment actually work?

Four steps. You apply on the labs platform and pick this track. You pay for the seat — a partner referral code comes off the price at that point. We review the application by hand. Once you are approved you get your labs account and the cohort schedule, and you start.

A partner gave me a referral code — where does it go?

If you arrived on a link carrying the code, it is already applied: you can see it on the price above, and it stays with you for thirty days even if you close the tab. If you have a code but no link, enter it at checkout. Either way the platform is the authority on what it is worth — the figure you see here comes from it, not from this page.

Is this a data-science course or an engineering one?

Engineering. The modelling is real — scikit-learn with proper cross-validation, PyTorch training loops, fine-tuning a transformer — but the destination is production: retrieval over a vector database, agents with tool use, experiment tracking, a model behind a FastAPI endpoint, and enough observability to know it still works next month. If the plan is to stay in a notebook, this is not the track.

Do you help with interviews and the job search?

That is a separate service, deliberately. The bootcamp buys you the skill and the work to show for it; our Career Solutions line covers resume and LinkedIn work, mock interviews and application support. Ask us and we will tell you honestly whether you need it yet.

Take the seat

The next batch is enrolling now. Three evenings a week, six weeks, and a place held for you as soon as your application is approved.

Enrol now → Talk to us first